doing 3D tracking based on an A-grid model output?

[keli559]

Learning how to do 3D tracking. However, the 3D tracking is based on a C-grid model output
Are there ways to do 3D tracking based on an A-grid model output? (meaning u v w are found at the same point)
thanks

[erikvansebille]

You can simply use the AdvectionRK4_3D Kernel. The only requirement is that your fieldset has a ‘W’ field, the grid type does not matter

[JamiePringle]

I know I am late in answering this -- but a word of caution. If you are using A grid data obtained by averaging a C-grid (e.g. the default Mercator Ocean output) you will find that the the A grid currents will run particles through land boundaries and the bottom. The land boundaries are only enforced on the C grid, not the averaged A grid. It is important to use the same grid as the model boundaries are implemented on. Copernicus will release C grid data if you ask them. Jamie

[erikvansebille]

Yes, very good point, Jamie! This issue with ill boundary conditions in interpolated fields is also what the Delandmeter and Van Sebille paper is (partially) about.

And note that the uninterpolated data for Copernicus is also available via the portal of Mercator Ocean. E.g. https://www.mercator-ocean.eu/en/solutions-expertise/accessing-digital-data/ and https://www.mercator-ocean.fr/wp-content/uploads/products_pdf/MOI_GLO12.pdf